Enhancing photoacoustic imaging with BFP-GAN

Abstract

This study introduces Binary Fourier Perception Generative Adversarial Network (BFP-GAN), a tailored approach for improved photoacoustic image generation in data-sparse environments. By combining GAN and Fourier decay perception techniques, the method markedly enhances image quality. Experimental validation using PAM and PAT datasets demonstrates its effectiveness in reducing artifacts and improving fidelity with limited data. BFP-GAN holds significant promise for diverse clinical applications, representing a major advancement in photoacoustic imaging technology.